This section covers the Segmentation Functions of the Document Imaging SDK. The result of the segmentation of an image is a same sized image, where all of the pixel values represent a region label. The region labels are numbers starting from 1. The image segmentation algorithms do not return a DIB like filter functions do, because of the large number of possible matrix values (region labels). Image segmentation algorithms returns WORD type matrices containing the output values. The matrices are allocated by the algorithms, and the user should use the implemented release functions to deallocate the memory. Due to the WORD type, the maximum number of the distinguishable regions is 65535. If the number of generatable regions is over the maximum, the regions labeled above the maximum will generate only one large region.